The present invention relates to a method for fabricating a semiconductor device; and, more particularly, to a method for fabricating a semiconductor device capable of reducing generations of seam when a self-aligned contact (SAC) plug is formed.
It is difficult to obtain process margins of a pattern formation process and overlay accuracy through a mere improvement on a level of integration in a semiconductor device. To solve these problems, a self-aligned contact (SAC) process is employed because it is cost-effective owing to a fact that an additional mask is not required for forming a contact hole pattern and the like. Among various schemes of carrying out the SAC process, the most typical scheme is to use a nitride layer as an etch barrier layer.
Also, because of the high level of integration, a contact process for forming an inter-layer contact, e.g., a plug, is employed. For instance, in about 0.15 xcexcm semiconductor devices, a hole-type contact mask is used in forming a bit line contact or a storage node contact. However, this use of the hole-type contact mask is not sufficient to secure a contact region due to a misalignment occurring during a photo-etching process. Therefore, a method of using different etch selectivity values between two different types of inter-layer insulation layers, e.g., an oxide layer and a nitride layer, is employed to secure the contact region. This method is employed in the aforementioned SAC process.
More specifically to the SAC process for forming a plug, an oxide layer for insulating a space between plugs is first etched to form a plug contact hole. Then, such material as polysilicon is deposited into the contact hole, and a chemical mechanical polishing (CMP) process is performed thereto to fill the polysilicon into the contact hole so that a plug is formed. Also, a T-type plug mask or an I-type plug mask is used in the SAC process for forming the plug.
However, in spite of the advantages of the SAC process, seams are more likely generated when the polysilicon is used as a plug material. The reason for this problem is because of a deterioration of topology caused by an undercut of an insulation layer. For example, the seam usually occurs at a storage node contact plug and a bit line contact plug when they are made of polysilicon.
Also, a chance of the seam generation is much higher at a portion of the insulation layer having a negative slope produced by the undercut of the insulation layer. Particularly, the seam is a main cause for degrading device characteristics. An increase of leakage currents is one example.
FIGS. 1A to 1E are cross-sectional views showing a conventional method for forming a SAC plug with use of the SAC process.
Referring to FIG. 1A, a plurality of device isolation layers 102 defining active regions 101 are formed in a substrate 100. A local oxidation of silicon (LOCOS) technique or a shallow trench isolation (STI) technique is employed for forming the device isolation layers 102. Also, each of the active regions 101 has an elongated elliptical shape when viewed from a top of the substrate 100. It should be also noted that there are a plurality of the active regions 101 defined by the device isolation layers 102 although they are expressed in a more simple representation for convenience.
A conductive layer 104A for forming a gate electrode (hereinafter referred to as a gate conductive layer) and a hard mask 104B for forming the gate electrode (hereinafter referred to as a gate hard mask) are sequentially formed on an entire surface of the substrate structure. Although it is not illustrated, an oxide-based insulation layer for forming the gate electrode (hereinafter referred to as a gate insulation layer) is formed beneath the gate conductive layer 104A. The gate insulation layer has a thickness in a range from about 50 xc3x85 to about 100 xc3x85. Herein, the gate conductive layer 104A is a single layer or a stacked layer of such materials as polysilicon, tungsten, tungsten nitride or/and tungsten silicide.
The gate hard mask 104B is made of such material like silicon nitride having a different etch selectivity from a subsequent inter-layer insulation layer 108 shown in FIG. 1B. Also, the gate hard mask 104B has a thickness ranging from about 1000 xc3x85 to about 2000 xc3x85.
For a lightly doped drain (LDD) structure, a low concentration of impurity ions for a source/drain is implanted into the active regions 101 formed at both sides of the gate electrode 104. Then, an etch stop layer 106 for forming a spacer for the gate electrode (hereinafter referred to as gate spacer) is deposited on the above entire substrate structure including the gate hard mask 104B and the gate conductive layer 104A. As like the gate hard mask 104B, the etch stop layer 106 is made of nitride having a different etch selectivity from the inter-layer insulation layer 108 shown in FIG. 1B. At this time, the etch stop layer 106 is deposited to a thickness in a range from about 300 xc3x85 to about 1000 xc3x85. However, it is much preferable to deposit the etch stop layer 106 to a thickness of about 500 xc3x85.
A photoresist pattern (not shown) is formed to make a core cell and a peripheral circuit regions opened. A blanket-etch process is then preformed to the etch stop layer 106 by using the photoresist pattern as an etch mask so that the gate spacer is formed at lateral sides of the gate electrode in the core cell and the peripheral circuit regions.
Next, a high concentration of impurity ions is implanted into the active regions 101 formed at both sides of the gate spacer to thereby form transistors in the core cell and the peripheral circuit regions. At this time, the etch stop layer 106 in the core cell array region is not etched to be used as another etch stop layer for the inter-layer insulation layer 108.
As shown in FIG. 1B, the inter-layer insulation layer 108 is formed on the above substrate structure including the etch stop layer 106. Herein, the inter-layer insulation layer 108 is made of an oxide layer having an excellent gap-fill property for preventing occurrences of a void phenomenon. Also, the inter-layer insulation layer 108 has a thickness ranging from about 3000 xc3x85 to about 9000 xc3x85. A preferable deposition thickness of the inter-layer insulation layer 108 is about 5000 xc3x85. Afterwards, a chemical mechanical polishing (CMP) process or a blanket-etch process is performed to planarize the inter-layer insulation layer 108. The planarized inter-layer insulation layer 108 remains on the gate hard mask 104B with a thickness T of about 1000 xc3x85.
A photoresist pattern 110 is formed on the inter-layer insulation layer 108 in such a manner that a region 111 for forming a SAC (hereinafter referred to as a SAC region) in the core cell array region is opened. The SAC region 111 can be a storage node contact region, a bit line contact region or a merged contact region obtained by merging the storage node contact region and the bit line contact region together. Herein, the illustrated SAC region is the merged contact region. The merged contact region is formed in a T-shape and includes a partial portion of the active region 101 and that of a non-active region.
In case of the T-shaped merged contact region, the size of the merged contact region is bigger than that of each storage node contact region and bit line contact region itself. As a result of this increased size, it is possible to prevent an etch-stop phenomenon usually occurring when the contact region is small. In addition, compared to a structure taught in an article by Kohyama et al. entitled xe2x80x9cA fully printable, self-aligned and planarized stacked capacitor DRAM cell technology for 1 Gbit DRAM and beyondxe2x80x9d, symp. On VLSI. Digest of Technical Papers, PP. 17-18, (1997), an occupying area of the photoresist pattern increases to thereby improve etch selectivity.
Next, the inter-layer insulation layer 108 and the etch stop layer 106 are sequentially etched until a partial portion of the active region 101 is exposed. From this etching, a plurality of contact holes 111A are formed. During the etching to the inter-layer insulation layer 108, the etch stop layer 106 serves,to play an etch-stop function.
Meanwhile, referring to FIG. 1C, the etching to the etch stop layer 106 results in a formation of a gate spacer 106A at lateral sides of the gate hard mask 104B in the core cell array region. Impurity ions are then implanted into the active regions 101 formed at both sides of the gate spacer 106A in order to reduce a contact resistance between a SAC plug and the active region 101 formed beneath the SAC plug.
As shown, after the photoresist pattern is removed 110, a conductive layer, for instance, a polysilicon layer 112 is deposited until completely being filled into the contact holes 111A. At this time, the deposition thickness ranges from about 3000 xc3x85 to about 7000 xc3x85. Afterwards, the polysilicon layer 112 is planarized by performing a CMP process or a blanket etch process until an upper surface of the inter-layer insulation layer 108 is exposed. In case of performing the CMP process to the polysilicon layer 112, slurry used for etching a typical polysilicon is employed.
Subsequently, a CMP process is performed again to the inter-layer insulation layer 108 and the polysilicon layer 112 until an upper portion of the etch stop layer 106 is exposed. This CMP process defines storage node contact plugs 112A and bit line contact plugs 112B electrically isolated from each other. The CMP process subjected to the inter-layer insulation layer 108 and the polysilicon layer 112 also employs slurry used for etching a typical oxide layer.
In case that the SAC contact region 111 as shown in FIG. 1B is not the merged contact region, the above CMP process for isolating electrically the storage node contact plugs 112A and the bit line contact plugs 112B can be omitted.
The seam is generated more frequently as an area of an opening portion of each contact hole 111A decreases. Particularly, the seam generation is more severe when an upper part of the gate electrode 104 gets to have a slope by etching the etch stop layer 106.
In addition to the above-described approach, another approach can be employed to secure sufficiently the contact region. First, an etching is stopped right above the etch stop layer, and a photoresist strip process and a wet cleaning/etching process are performed thereafter. Then, in the step of removing the etch stop layer, a capping layer is deposited on the etch stop layer to secure a required thickness of the gate hard mask. Herein, the capping layer is made of such material having a poor coverage property as plasma enhanced chemical vapor deposition (PECVD) oxide or undoped silicate glass (USG). After the deposition of the capping layer, the oxide layer gets to remain only on the gate hard mask through the use of a wet cleaning/etching process, and the nitride layer is then removed through the use of a dry etching process.
However, this approach results in an undercut of the inter-layer insulation layer during the wet cleaning/etching process. This undercut further induces the seam generations when the SAC plug is subsequently formed. Also, the seams are more severely generated as an executing period of the wet cleaning/etching process for extending the contact hole region is longer.
FIG. 2 is a top-view of the typical semiconductor device completed with the SAC process for forming the plug. As shown, a plurality of device isolation layers 200 are allocated on a substrate structure 200. A plurality of gate electrodes are arrayed in a direction of crossing the device isolation layers 201. Herein, the reference numeral 203 is a region opened for forming a plug, i.e., a contact hole.
FIGS. 3A and 3B are cross-sectional views of FIG. 4 in each direction of the lines A-Axe2x80x2 and B-Bxe2x80x2.
Referring to FIG. 3A, a plurality of device isolation layers 201 are formed in a substrate 200, and then, an oxide-based gate insulation layer 202A, a gate conductive layer 202B and a gate hard mask 202C are sequentially deposited on the substrate structure. Afterwards, a photo-etching process is performed with use of the gate hard mask 202C to form a gate electrode 202. Herein, the gate conductive layer 202B is a single layer or a stacked layer of tungsten, polysilicon or tungsten silicide. Also, the gate hard mask 202C is a nitride-based layer such as a silicon nitride layer or a silicon oxynitride layer.
Next, an etch stop layer 202D made of silicon nitride or silicon oxynitride is formed at lateral sides of the gate electrode 202. An inter-layer insulation layer 204 is then formed in such a manner to be filled into a space between the gate electrodes 202. At this time, the inter-layer insulation layer 204 is preferably formed to a thickness ranging form about 2000 xc3x85 to about 10000 xc3x85 by using a material having a good planarization property such as high temperature oxide (HTO), advanced planarization layer (APL) oxide, spin on dielectric (SOD), spin on glass (SOG), tetra-ethyl-ortho silicate (TEOS), boro-phospho-silicate glass (BPS G), phospho-silicate glass (PSG) or boro-silicate glass (BSG). Also, it is preferable to perform a deposition or a deposition/planarization process so that a thickness of the inter-layer insulation layer 204 on an upper surface of the gate hard mask 202D ranges from about 0 xc3x85 to about 1000 xc3x85.
Next, a plurality of contact holes 203 for bit line contacts or storage node contacts are formed. More specifically, a photoresist pattern (not shown) for forming the contact holes 203 is formed, and an upper part of an impurity diffusion region (not shown) allocated between the gate electrodes 202 is opened through the use of the typical SAC process.
In more detail of the SAC process, it is possible to use different types of etch gas to attain different effects. When the inter-layer insulation layer 204 made of BPSG and the like is etched, an etch gas containing carbons and inducing lots of polymers is used to provide a high etch selectivity with respect to the nitride-based layers, i.e., the gate hard mask 202C and the etch stop layer 202D. Examples of this etch gas are C3F8, C4F8, C5F8, C4F6 and C2F4. Also, such gas as CHF3, C2HF5, CH2F2, CH3F, CH2, CH4, C2H4 and H2 can be also used to provide a reliable etch process by increasing an etch process margin along with the high etch selectivity. Also, such inert gas as He, Ne, Ar, Kr or Xe can be also used as the etch gas to improve an etch stop function by enhancing a sputtering effect and plasma stability. It is also possible to use a gas obtained by mixing the above etch gases with each other. It is further possible to add CxHyFz, where x, y and z is equal to or greater than 2, to the etch gas containing lots of carbons to secure margins of the etch process.
As shown in FIG. 3B, another insulation layer is deposited on the inter-layer insulation layer 204 and the gate electrode 202 to form a capping layer 205 with an over-hang structure. Herein, the capping layer 205 is made of USG having a less powerful coverage property, and its function is to prevent losses of the gate hard mask 202C during a removal of the etch stop layer 202D for exposing a surface of the substrate 206.
However, the capping layer 205 causes degradation of a gap-fill property, and this degradation further induces generations of void and seam in case that a conductive layer for forming a plug (hereinafter referred to as plug conductive layer) is deposited after the contact holes 203 are extended through a subsequent wet cleaning/etching process and the surface of the substrate 200 is then exposed through a blanket-etch process.
Also, the generations of void and seam are caused by a bowing profile phenomenon, wherein a profile obtained prior to depositing a plug material is bent. Particularly, the capping layer 205, the etch stop layer 202D and the wet cleaning/etching process used for extending the contact holes 203 are adopted to meet trends of a decrease in pattern size and an increase in a difference in height. However, these implementations of the capping layer 205, the etch stop layer 202D and the wet cleaning/etching process become a factor for causing the bowing profile phenomenon.
The USG typically used for the capping layer 205 has a slower wet etch rate than that of BPSG typically used for the insulation layer. This usage of the USG material results in the bowing phenomenon. Particularly, the bowing phenomenon becomes more severe as an executing period of the wet cleaning/etching process for extending an opening portion of the contact hole is longer. Additionally, instead of the USG, plasma enhanced tetra-ethyl-ortho silicate (PETEOS) having a poor coverage property can be also adopted in the capping layer 205 with the over-hang structure.
As described above, the generation of the void and seam is resulted from a negative slope of the insulation layer produced by the undercut of the insulation layer. After the step of isolating the plugs, this undercut of the insulation layer further pronounces the seam generations at the plug. The seam generated at the plug becomes a main factor for degrading device characteristics. For instance, leakage currents are increased.
It is, therefore, an object of the present invention to provide a method for fabricating a semiconductor device capable of effectively preventing generations of void and seam at a plug due to a negative slope of an insulation layer and a bowing phenomenon in an etch profile of a contact hole.
In accordance with an aspect of the present invention, there is provided a method for fabricating a semiconductor device, including the steps of: forming a plurality of conductive patterns on a substrate; forming an etch stop layer along the plurality of the conductive patterns; forming an insulation layer on an entire surface of the substrate structure; etching selectively the insulation layer to form a plurality of contact holes exposing a portion of the etch stop. layer allocated in between the conductive patterns; forming an attack barrier layer for preventing the insulation layer from being attacked by a chemical used in a wet cleaning/etching process along a profile containing the contact hole; forming a capping layer having an over-hang structure on an upper part of each conductive pattern; extending an opening portion of each contact hole by performing a wet cleaning/etching process to a bottom side of each contact hole; removing selectively a portion of the etch stop layer and the attack barrier layer disposed at the bottom side of each contact hole to expose a surface of the substrate; and forming a plug contacted to the exposed surface of the contact hole.
In accordance with another aspect of the present invention, there is also provided a method for fabricating a semiconductor device, including the steps of: forming a plurality of conductive patterns on a substrate; forming an etch stop layer along the plurality of the conductive patterns; forming an insulation layer on an entire surface of the substrate structure; etching selectively the insulation layer to form a plurality of contact holes exposing a portion of the etch stop layer allocated in between the conductive patterns; forming a capping layer having an over-hang structure on an upper part of each conductive pattern; weakening bonding forces between atoms contained in sidewalls of the capping layer with use of an inert gas; extending an opening portion of the contact hole by performing a wet cleaning/etching process and simultaneously removing the sidewalls of the capping layer; removing selectively a portion of the etch stop layer disposed at a bottom side of each contact hole to expose a surface of the substrate; and forming a plug contacted to the exposed surface of the substrate.